Supply chain · Deep dive

Altana

A federated AI map of the global supply chain — enterprises and governments pool proprietary trade data without exposing it, and Altana sells the resulting value-chain 'source of truth' as compliance, risk, and tariff intelligence.

emerging

The question that decides it: Altana's edge is a federated network where enterprises and governments contribute proprietary supply-chain data without exposing it, making a value-chain map the company claims is twice as rich as anyone else's. Does that network compound into a real data moat — each new participant making the map materially better and harder to copy — or do commoditized bill-of-lading data plus generative AI let Sayari, Kpler, or an S&P/Panjiva incumbent rebuild a 'good enough' map without the federation, collapsing Altana's differentiation to its government relationships?

My take ★★ Very interesting ★ Interesting − Maybe ✗ Not interesting

Snapshot

Altana is trying to build the single trusted map of the physical global economy — who makes what, where, for whom, and through which suppliers — and to do it without any company or government having to hand over its confidential data. It stitches together public customs and bill-of-lading records with the private, first-party supplier data of its customers using federated machine learning, so participants share intelligence without sharing the underlying data. On top of that map it sells compliance, supplier risk, tariff modeling and trade-enforcement tools to both multinationals (Maersk, Boston Scientific, L.L.Bean) and governments (U.S. Customs and Border Protection, the UK Department for Business and Trade). Founded in 2018 by three Panjiva alumni, Altana crossed into unicorn territory in July 2024 with a $200M Series C at a $1B valuation, bringing total funding to roughly $322M. The timing is its tailwind: forced-labor enforcement, tariff whiplash and reshoring have made value-chain visibility a board-level problem.

Founding story

The company is a direct descendant of Panjiva, the global trade-data pioneer founded in 2006 and acquired by S&P Global in February 2018. Evan Smith, Peter Swartz and Raphael Tehranian all worked there — Smith on enterprise solutions and partnerships, Swartz as head of data science, Tehranian on the commercial side — and left after the S&P deal to found Altana in December 2018. Their frustration was specific: Panjiva and its peers had spent a decade digesting customs and shipment records, but that data only ever describes the transactions crossing a border. It cannot see inside a company’s own supplier relationships, and no enterprise or government will pour its confidential supplier lists into a shared database it does not control.

Their bet was that the missing piece was not more public data but a way to fuse public data with everyone’s private data without anyone giving up custody of it. Federated learning — training shared models across many parties’ data without moving that data — was the mechanism. Smith had earlier run IMBU Technologies, a textile supply-chain automation firm, so he had felt the gap between raw trade data and operational supplier management first-hand. Amadeus Capital Partners, a firm with a long history in AI, risk and national-security software, became the first institutional backer with a $7M seed in November 2020, joined by Schematic Ventures, AlleyCorp and the Working Capital supply-chain fund. GV (Google Ventures) led the $15M Series A the following year.

How it works

Start with the base layer: Altana ingests the world’s public trade exhaust — customs filings, bills of lading, shipment manifests — and uses AI to resolve messy records into a knowledge graph. It identifies the real shipper and consignee behind aliases, classifies what is actually inside a container, links entities across languages and jurisdictions, and infers multi-tier relationships (your supplier’s supplier’s supplier). That alone produces a global map of companies and the goods moving between them.

The federated layer is what the company claims makes the map “more than twice as rich” as anyone else’s. When a customer joins, it connects its own siloed data — product catalogs, supplier master lists, purchase orders, factory relationships — to its private instance of the map. Altana’s models learn from that first-party data to improve the shared map’s accuracy and completeness, but the raw customer data never leaves the customer’s control and is never exposed to other participants. In effect, each new enterprise or government makes the collective map sharper for everyone while keeping its own IP and sovereignty. On top of that map sit AI agents: an “agentic” workflow can automatically assign Harmonized System (HS) codes to goods, calculate country of origin under trade rules, audit shipments, and run continuously as a kind of always-on compliance officer that surfaces only exceptions and risks. The Maersk partnership extends this into “Product Passports,” where AI-suggested product attributes travel with the goods across ports.

Product and business overview

The core product is Altana Atlas — the value-chain map and the applications built on it, marketed since June 2024 as a “Value Chain Management System.” The applications cluster into a few named jobs. Supply chain visibility and mapping surfaces multi-tier suppliers a company did not know it depended on. Compliance and risk screens the network for forced-labor exposure (critical for the U.S. Uyghur Forced Labor Prevention Act), sanctioned entities, and other regulatory flags. Procurement and sourcing helps design resilient supplier lines and find alternatives. The Tariff Scenario Planner — heavily promoted through 2025-2026 amid U.S. tariff volatility — lets companies model duty impacts across their entire extended supplier network and re-source to minimize exposure. For governments, the same map powers trade enforcement, counter-narcotics (CBP is using it to investigate fentanyl precursor networks), and traceability programs.

The dual customer base is the strategic core. Enterprises pay for visibility, compliance and cost; governments pay for enforcement and national security. Because both feed the federated map, Altana’s pitch is that its public and private customers make each other’s product better — a network effect that a pure data vendor or a pure SaaS tool cannot reproduce.

Business model and pricing

Altana sells enterprise SaaS: annual or multi-year subscriptions to Atlas and its modules, priced by deployment complexity, supplier network size, and which applications are switched on. There is no public rate card — deals are custom, and the enterprise tier is explicitly aimed at governments, large multinationals and complex logistics networks with 10,000-plus Tier-1 suppliers, network-wide risk automation, and bespoke AI models with dedicated onboarding. Third-party software listings describe it as a quote-only enterprise product, which is consistent with a land-and-expand motion: start with one use case (say UFLPA compliance), then expand into procurement, tariffs and trusted-trade programs across the same account.

Government revenue is more visible because contracts leave a paper trail. Altana’s initial CBP/DHS engagement, disclosed in 2023, was worth about $2.85M in year one and close to $10M over three years; in October 2025 CBP signed a formal two-year contract to run Altana’s platform for real-time trade enforcement, forced-labor and counter-narcotics detection, and customs modernization, alongside the Product Passports program. Those are meaningful reference accounts, but they are also lumpy, procurement-cycle-dependent revenue — a different animal from smooth enterprise ARR, and a reason to treat any single-year revenue figure cautiously.

Traction over time

The direction is clearly up and to the right: from a $7M seed in 2020 to a $1B mark in mid-2024, with GV, Activate Capital, Thomas Tull’s US Innovative Technology Fund, Generation Investment Management, March Capital and Salesforce Ventures on the cap table. Named customers grew to include Maersk, Boston Scientific, General Atomics, L.L.Bean, CBP and the UK Department for Business and Trade. Two honest caveats. First, revenue is undisclosed; the only number in circulation is getlatka’s ~$37.5M ARR estimate for 2024, which is unverified and, if roughly right, implies the $1B mark was struck at something like 25-30x revenue — a rich multiple that prices in the network-effect thesis, not current cash flows. Second, there is no publicly reported Series D as of mid-2026, so the $1B valuation is now two years stale in a market where AI-adjacent supply-chain multiples have both soared and reset.

Market analysis

The addressable market depends heavily on how you draw the box. Supply-chain risk-management software alone was sized anywhere from ~$3.5B to ~$9B in 2025 across research houses (Market Research Future, Mordor, SkyQuest), growing 10-21% a year depending on definition. But Altana sits at the intersection of several budgets — trade compliance, supplier risk, procurement, ESG/forced-labor, and government enforcement — so the real pool it hunts is larger and more fragmented than any single “SCRM software” figure. The structural forces are unusually favorable: the UFLPA and similar forced-labor rules made supply-chain traceability legally mandatory; the 2025-2026 tariff environment turned country-of-origin and duty modeling into a CFO-level cost problem (Altana said its tariff calculator usage spiked 213% in a single week); and reshoring plus deglobalization are pushing companies to re-map suppliers they never scrutinized. The counterforce is that “visibility” is a crowded, partly commoditized category, and much of the underlying data is cheaply buyable.

Competitive intel

Altana is boxed in on several sides. Sayari is the sharpest direct rival — corporate-ownership and supply-chain risk intelligence for governments and compliance teams, now PE-backed after TPG’s ~$235M majority investment in April 2024, and chasing the same national-security accounts. Exiger competes head-on for federal compliance, UFLPA and defense-industrial-base contracts — the exact lane where Altana won CBP. Interos ($1B+ valuation in 2021) sells a similar multi-tier risk-network graph to enterprises. Kpler has rolled up trade-data assets (ImportGenius, MarineTraffic) and owns maritime/commodity flow analytics at scale. And looming behind all of them is S&P Global Market Intelligence / Panjiva — the incumbent the founders left, which owns the customs-data layer and has distribution into essentially every large enterprise and government desk. Altana’s real differentiation is the federated first-party data architecture: rivals mostly analyze public data, while Altana claims to fuse public plus private without exposing the private. If that federation genuinely compounds, it is a moat. If generative AI plus cheap bill-of-lading data lets a rival build a “good enough” map without the federation, the differentiation thins to Altana’s government relationships — which are valuable but replicable by the likes of Sayari and Exiger.

History and evolution

• Feb 2018 — S&P Global acquires Panjiva, the trade-data firm where all three founders work.
• Dec 2018 — Smith, Swartz and Tehranian found Altana in Brooklyn to build a federated value-chain map.
• Nov 2020 — $7M seed led by Amadeus Capital Partners; first institutional capital.
• Sep 2021 — $15M Series A led by GV (Google Ventures).
• Oct 2022 — $100M Series B led by Activate Capital, with strategic investors Prologis and Reefknot; big step-up in ambition and headcount.
• 2023 — Wins DHS/CBP contract (~$2.85M year one, ~$10M over three years) to support UFLPA forced-labor enforcement.
• Jun 2024 — Launches its “Value Chain Management System” framing for Atlas.
• Jul 2024 — $200M Series C led by Thomas Tull’s US Innovative Technology Fund at a $1B valuation; joins the unicorn ranks.
• 2025 — Heavy push on the Tariff Scenario Planner amid U.S. tariff upheaval; expands UK government work; partners with Maersk.
• Oct 2025 — CBP signs a two-year contract and selects Altana’s AI-powered Product Passports for supply-chain traceability, deployed with Maersk across ports handling a large share of global trade.

What people say

The case for. Employees on Glassdoor rate Altana around 4.0/5 with roughly 74% recommending it (2026), praising high talent density, working on genuinely hard problems, and marquee customers. Investors clearly buy the network thesis: GV, Generation IM, March Capital, Salesforce Ventures and Thomas Tull’s fund backed a $1B mark, and Generation’s public write-up framed Altana as the way to bring AI to supplier intelligence. The product wins where the pain is acute — the CBP, UK government and Maersk relationships are real, dated, and hard to fake, and the tariff-planner usage spike shows the map converts to demand when trade rules move. The federated architecture is a genuinely differentiated answer to the real objection (“I won’t share my supplier data”), which is why it keeps winning sovereignty-sensitive government work.

The complaints. The Glassdoor picture has a sharp underside: multiple reviews describe a culture of fear, frequent and poorly explained layoffs (“a bonus one week, layoffs the next”), teams stretched thin with undersized HR support, and specific criticism that the company is dismissive of women’s contributions despite marketing diversity as a value. The recurring theme is instability and management turnover — normal for a fast-scaling startup, but pointed here. On the business side, the honest risks are structural: revenue is undisclosed and the only ARR estimate (~$37.5M for 2024) implies a very rich multiple; government revenue is lumpy and procurement-dependent; the underlying trade data is partly commoditized; and the whole differentiation rests on the federated network actually compounding rather than being matched by cheaper AI-plus-public-data rivals. No Series D since 2024 leaves the $1B mark unrefreshed while competitors like Sayari take on fresh PE capital.

Outlook: the open question

Altana has built something genuinely novel — a way to pool the world’s supply-chain data without anyone surrendering custody of it — and it has aimed that architecture at exactly the problems (forced-labor enforcement, tariff modeling, trade security) that regulation and geopolitics have made unavoidable. The government wins are real and the enterprise logos are credible. But the $1B mark, struck in July 2024 against perhaps ~$37.5M of estimated ARR, is a bet on the network, not the current business. Altana becomes a durable, category-defining company if the federated data network compounds into a true moat — each new enterprise and government participant making the map measurably richer and harder to replicate, driving expansion revenue across compliance, procurement and tariffs, and turning lumpy government contracts into a reference base that pulls in commercial ARR faster than headcount and burn. It stalls, and gets repriced, if the differentiation proves thinner than claimed — if commoditized bill-of-lading data plus generative AI let Sayari, Exiger, Kpler or an S&P/Panjiva incumbent stand up a “good enough” value-chain map without the federation, collapsing Altana’s edge to its government relationships (valuable, but winnable by rivals) while a rich private mark waits for a down-round refresh. The evidence to watch: net revenue retention and multi-module expansion inside enterprise accounts, whether the tariff-planner surge converts to recurring contracts, the terms of the next raise, and whether the internal churn the reviews describe stabilizes as the company scales. The mechanism, not the mission, decides this one.

How a challenger would attack it

Rebuild the map without the federation. Altana’s differentiation rests on the claim that federated first-party data makes its map “more than twice as rich” as public-data rivals. The attack is to test that claim with cheap inputs: bill-of-lading and customs records are commoditized — ImportGenius sells them across 14+ countries — and modern entity-resolution models make the graph-stitching that took Altana years steadily cheaper to replicate. A challenger ships a “good enough” map at a transparent price against Altana’s quote-only enterprise tier, and wins the mid-market Altana’s 10,000-Tier-1-supplier deployment model ignores. The second vector is speed on the wedge use case: the tariff-planner usage spike (213% in a week) shows demand arrives in bursts when trade rules move — a product-led, self-serve tariff and origin calculator captures that surge while Altana runs bespoke onboarding. Third, contest the government lane directly, as Sayari (fresh TPG capital) and Exiger already do: CBP-style contracts are procurement-cycle-dependent and re-compete on a two-year clock, and Altana’s Glassdoor-documented churn and layoff cycles create hiring openings for exactly the government-facing talent those bids need. The unrefreshed 2024 $1B mark against ~$37.5M estimated ARR means Altana can’t easily raise to out-spend a well-capitalized attacker.

Same playbook, new buyer

The federated architecture — pool proprietary data across parties who refuse to share it — is the reusable invention, and supply chain is only one place where that objection blocks value. The nearest shift is financial crime and trade finance: banks sit on transaction and counterparty data they legally cannot pool, yet all need the same network view for sanctions and TBML screening; a federated map sold to bank consortia runs Altana’s playbook for a buyer with far bigger compliance budgets. Within trade itself, the geographic shift is the EU: CBAM, the EU forced-labor regulation and deforestation rules create UFLPA-style mandatory traceability, but Altana’s anchor relationships (CBP, UK DBT) and its US Innovative Technology Fund lead investor stamp it as an American national-security asset — a real handicap when selling data sovereignty to Brussels or to Asian governments wary of US-aligned platforms. A neutral, EU-domiciled federation wins there almost by default. Down-market is the third opening: mid-size importers need UFLPA screening and origin calculation but will never fund a bespoke AI deployment; Altana’s land-and-expand enterprise motion and government-grade cost structure make it structurally unable to chase $20K ACV self-serve accounts.

Sources and further reading

• Altana, An AI Platform For Supply Chain Data, Hits Unicorn Status After $200 Million Raise (Forbes, Jul 2024)
• Supply-chain startup Altana valued at $1B in Thomas Tull’s latest AI bet (PitchBook, Jul 2024)
• AI-powered supply chain visibility platform Altana bags $100M (TechCrunch, Oct 2022)
• Altana raises $7M to protect supply chains from disruptions, child labor (TechCrunch, Nov 2020)
• Altana Introduces World’s First Value Chain Management System (BusinessWire, Jun 2024)
• Exclusive: CBP taps Altana’s AI for trade enforcement (Axios, Oct 2025)
• How one AI company is helping businesses navigate Trump’s tariff chaos (Fortune, Feb 2026)
• TPG Announces $228 Million Strategic Growth Investment in Sayari (TPG, Jan 2024)
• Altana AI Reviews — employee ratings (Glassdoor, accessed 2026)
• Altana: from Seed to Unicorn with Amadeus (Amadeus Capital, 2024)